CN113609782B - Real-time prediction method and simulation system for wave force applied to motion load - Google Patents

Real-time prediction method and simulation system for wave force applied to motion load Download PDF

Info

Publication number
CN113609782B
CN113609782B CN202110943138.5A CN202110943138A CN113609782B CN 113609782 B CN113609782 B CN 113609782B CN 202110943138 A CN202110943138 A CN 202110943138A CN 113609782 B CN113609782 B CN 113609782B
Authority
CN
China
Prior art keywords
wave
pool
numerical
grid
floating body
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110943138.5A
Other languages
Chinese (zh)
Other versions
CN113609782A (en
Inventor
马昕
王凯
宋锐
荣学文
李贻斌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong University
Original Assignee
Shandong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong University filed Critical Shandong University
Priority to CN202110943138.5A priority Critical patent/CN113609782B/en
Publication of CN113609782A publication Critical patent/CN113609782A/en
Application granted granted Critical
Publication of CN113609782B publication Critical patent/CN113609782B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/08—Fluids
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Algebra (AREA)
  • Computing Systems (AREA)
  • Fluid Mechanics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention provides a real-time prediction method and a simulation system for motion load wave force, comprising the following steps: establishing a numerical wave pool and a solver; the interaction of the floating bodies in two-dimensional and three-dimensional different shapes and the waves with different parameters is solved by using the solver, and the motion gesture of the floating bodies along with the waves and the stress information at each moment are calculated; and training the neural network by using the calculated motion gesture and stress information, and predicting the wave force of the floating body in a specific environment on line by using the trained neural network. According to the invention, a controller can be designed in the numerical wave pool, a controller interface is preset in the numerical wave pool, namely, a control equation can be added in a configuration file, in the simulation calculation process, the control force is obtained through feedback information calculation and then applied to a controlled object, and the object is controlled to stabilize the object motion. The numerical wave pool can conveniently simulate the motion condition of floating bodies of different objects under different sea conditions, and has great flexibility and practicability.

Description

Real-time prediction method and simulation system for wave force applied to motion load
Technical Field
The invention belongs to the technical field of real-time prediction of wave force, and particularly relates to a real-time prediction method and a simulation system of wave force applied to a motion load.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Under the condition of severe sea conditions, the construction of safely and stably hanging goods into sea water for engineering is an extremely important problem in ocean engineering.
Because the deep sea crane is fixed on a moving platform such as a ship, the ship can generate movements such as heave and roll under the influence of sea waves. When the suspended load contacts seawater, the load is also influenced by waves to generate nonlinear irregular movements, and the movements can cause collision of the load and a ship body, so that damage to the load or breakage of a cable is caused, further serious accidents are caused, and life and property safety is damaged. This requires the control system of the deep sea crane to be designed so that the work therein is performed safely and efficiently.
The inventor finds that the time lag phenomenon is a phenomenon frequently occurring in practical engineering, and the network transmission speed and the longer sampling time of the sensor can cause the time lag phenomenon of data feedback, so that the timely obtaining of the motion and stress data of the ship body and the load in water is difficult, and particularly when the load is far away from the crane, the stress information of the load cannot be timely fed back into a control system of the crane. These factors all increase the difficulty of designing the deep sea crane controller.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a real-time prediction method of the motion load wave force, which can provide the stress information of the floating body (ship and load) on line.
To achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
in a first aspect, a method for real-time prediction of wave forces on a moving load is disclosed, comprising:
establishing a numerical wave pool simulating the interaction of the simulated wave and the floating body and a solver solving the interaction of the wave and the floating body;
the interaction of the floating bodies in two-dimensional and three-dimensional different shapes and the waves with different parameters is solved by using the solver, and the motion gesture of the floating bodies along with the waves and the stress information at each moment are calculated;
and training the neural network by using the calculated motion gesture and stress information, and predicting the wave force of the floating body in a specific environment on line by using the trained neural network.
According to a further technical scheme, the numerical wave pool is a three-dimensional simulation model, an origin o is placed at the bottom of the numerical wave pool, the parallel direction of the x axis and the still water surface is horizontal to the right, the z axis is vertical upwards, the y axis is the direction in which the width edge of the numerical wave pool is located, and waves propagate along the x axis direction.
According to a further technical scheme, the numerical wave pool grid and boundary condition setting are required to be set before the numerical wave pool simulates interaction between the simulated waves and the floating body.
Further preferably, the set value wave pool grid is specifically:
setting the size of a numerical wave pool;
and integrating each grid by using the hexahedral grid discrete calculation region and obtaining a discrete equation set of the whole calculation region.
Further preferably, when the hexahedral mesh is used for discrete calculation of the area, the discrete mesh is set to be a non-uniform or uniform-size mesh, and the variation degree of the flow field variable in the calculation area is positively correlated with the density degree of the mesh;
in the simulation of wave and object interactions, the wave generation boundary, i.e., the left side region of the pool, the floating body region, i.e., the middle region of the pool, and the wave absorption region, i.e., the right side region of the pool, employ higher density grids.
Further technical proposal, when boundary conditions are set, the constrained variables comprise speed, pressure and free surface;
the numerical wave pool adopts a fluid volume method to solve the shape of the free liquid level, the phase fraction alpha is used for distinguishing air from liquid, the grid cell is liquid when the phase fraction alpha is 1, the grid cell is filled with air when the phase fraction alpha is 0, and the grid cell exists at the free liquid level when the phase fraction alpha is between 0 and 1;
preferably, the side of the numerical wave pool is a wall type boundary, the water is prevented from flowing out, the top is an atm sphere boundary, and the backflow phenomenon is prevented, so that the pressure in the pool is balanced.
According to a further technical scheme, the floating body is generated and then is led into a numerical wave pool, and a required grid is generated around the floating body.
According to a further technical scheme, the trained neural network predicts the wave force of the floating body in a specific environment on line, and specifically comprises the following steps:
each time step, the neural network receives a group of load gesture data, and the two layers of GRU units expand input information to a high-dimensional space for processing;
the GRU is a cyclic neural network, and hidden information of the input data is recorded through the memory unit h;
and finally, the data passes through a full connection layer to output stress information of the object.
According to the technical scheme, the input of each time step of the neural network is a one-dimensional vector with a plurality of elements, the one-dimensional vector comprises wave parameters and attitude information of an object, and the output of the one-dimensional vector is the force and moment born by the floating body in the directions of x, y and z axes.
In a second aspect, a real-time prediction system for a moving load subjected to wave forces is disclosed, comprising:
a simulation model building module configured to: establishing a numerical wave pool simulating the interaction of the simulated wave and the floating body and a solver solving the interaction of the wave and the floating body;
a solution module configured to: the interaction of the floating bodies in two-dimensional and three-dimensional different shapes and the waves with different parameters is solved by using the solver, and the motion gesture of the floating bodies along with the waves and the stress information at each moment are calculated;
an online prediction module configured to: and training the neural network by using the calculated motion gesture and stress information, and predicting the wave force of the floating body in a specific environment on line by using the trained neural network.
The one or more of the above technical solutions have the following beneficial effects:
the numerical pool based on computational fluid dynamics is used for calculating the interaction between the floating body and waves, a large amount of data are generated by the numerical pool to train the neural network, the trained neural network is used for providing stress information of the floating body (ship and load) for a crane control system on line, the needed stress information can be obtained on line by adopting the mode, and a controller can conveniently make a control strategy according to the stress information, so that accurate floating body control is realized.
The numerical wave pool disclosed by the invention can be used for conveniently simulating the motion conditions of floating bodies of different objects under different sea conditions, is lower in cost, and is more flexible in model establishment and setting, so that the numerical wave pool has great flexibility and wide practicability. The neural network based on the GRU structural design is used for rapidly predicting the load stress of the load in the waves in real time.
The invention uses the method of computational fluid mechanics to obtain the wave force suffered by the moving floating body under different parameter waves through the numerical wave pool, uses the force data as a training set to train the neural network, the trained neural network can predict the wave force suffered by a specific floating body under a specific environment, the trained neural network can be used for the design of the controller of the deep sea crane, and the trained neural network can be arranged in the computational fluid mechanics software OpenFOAM to test the performance of the controller of the deep sea crane.
In addition, the numerical wave pool based on computational fluid dynamics can be also used for verifying the performance of the designed deep sea crane controller.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is a diagram of a numerical wave pool model in accordance with an embodiment of the present invention;
FIG. 2 is a front view of a numerical wave pool grid distribution in accordance with an embodiment of the present invention;
FIG. 3 is a grid view of a "convex" shaped floating body in accordance with an embodiment of the present invention;
FIG. 4 (a) is a front view of a computing domain and a corresponding parameter setting diagram according to an embodiment of the present invention;
fig. 4 (b) is a schematic diagram of verification of the load rotation amount and heave displacement result obtained by the current model and the physical experiment according to the embodiment of the present invention;
FIG. 5 is a schematic diagram of a network model for predicting wave stress according to an embodiment of the present invention;
FIG. 6 is a diagram of training set data and predictive value mean square values according to an embodiment of the invention;
FIG. 7 is a diagram showing comparison between test data and reference data in case A1 of the present invention;
FIG. 8 is a diagram showing comparison between test data and reference data in case D1 according to the embodiment of the present invention;
FIG. 9 is a diagram showing comparison between test data and reference data in case T1 according to the embodiment of the present invention;
FIG. 10 is a diagram illustrating the mean square difference between test set data and predicted values according to an embodiment of the present invention;
FIG. 11 is a diagram showing comparison between test data and reference data in case C1 of the present invention;
FIG. 12 is a diagram showing comparison between test data and reference data in case C2 of the present invention;
FIG. 13 is a schematic diagram of a controller stabilized float designed in a numerical wave pool according to an embodiment of the present invention.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present invention.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
The embodiment discloses a real-time prediction method of the wave force of a motion load, which combines a method of computational fluid dynamics software OpenFOAM and GRU neural network to predict the wave force of the motion load in real time.
The numerical pool based on computational fluid dynamics is used for calculating interaction between the floating body and waves, a large amount of data is generated by the numerical pool to train the neural network, and the trained neural network is used for providing stress information of the floating body (ship and load) for a crane control system on line. And the numerical wave pool based on computational fluid dynamics can be also used for verifying the performance of the designed deep sea crane controller.
In an embodiment, the trained neural network may be introduced into the OpenFOAM to predict data by introducing a pyresch library, so as to expand the functions of the OpenFOAM.
Specifically, the method is divided into two parts, wherein the first part is that firstly, a numerical wave pool for simulating the interaction between the simulation wave and the floating body is established in open source computational fluid dynamics software OpenFOAM and a waves2Foam library, and on the basis of the waves2Foam self-contained solver, the dynamic grid solving and updating process of the inter DyMFOAM is added into the waves2Foam, so that the object motion problem can be processed by utilizing the dynamic grid technology in the process of solving the two-phase flow problem, and a solver suitable for solving the interaction between the wave and the motion object is established. The numerical wave pool comprises: a meshing tool, a wave generator, an initial boundary condition configuration tool, an object mesh generation tool, and a solver. The numerical wave pool and the newly established solver can solve the interaction of two-dimensional and three-dimensional objects in different shapes and waves with different parameters, and can calculate and obtain the motion gesture of the objects along with the waves and the stress information at each moment.
In a specific implementation example, a numerical wave pool based on a combination of computational fluid dynamics software OpenFOAM and waves2Foam tools is first used.
The numerical pool model is shown in fig. 1, and the simulation model is three-dimensional, and preferably, the simulation model is combined with a waves2Foam tool, the boundary of the pool inlet can be customized to various types of waves, and loose areas can be arranged at the inlet and the outlet to eliminate unnecessary wave reflection, and the object motion process is processed through a six-degree-of-freedom rigid body solver and a deformed grid technology. The numerical model is also provided with a controller design interface, a control equation of the control system can be added into the numerical model at will, and control force is applied to the controlled object in the solving process. The coordinate system (o-xyz) is set as shown in fig. 1, the origin o is placed at the bottom of the numerical pool, the x-axis and the hydrostatic surface are parallel to each other and horizontally right, the z-axis is vertically upward, the y-axis is the direction in which the width edge of the numerical wave pool is located, and the waves propagate along the x-axis direction.
Then, the simulated wave interacts with the object, including:
step one: numerical value wave pool grid setting:
the size of a numerical pool is set through a block mesh tool of the OpenFOAM, a hexahedral grid discrete calculation area is utilized, and a control equation is integrated on each grid to obtain a discrete equation set of the whole calculation area. The grid used for dispersion can be set to be uneven, and the area with larger flow field variation can be dispersed by using a denser grid to improve the accuracy of calculation. Grid distribution front view as shown in fig. 2, in the simulation of wave-object interaction, the wave generation boundary (i.e., the left side area of the pool), the area where the floating body is located (the middle area of the pool), and the wave-absorbing area (i.e., the right side area of the pool) employ grids with higher density.
Step two: boundary condition setting:
the boundary conditions of the numerical pool of waves have a constraining effect on the field of the numerical pool. The constrained variables include velocity, pressure, and free surface. The numerical pool adopts a fluid volume method to solve the shape of the free liquid level, the phase fraction alpha is used for distinguishing air from liquid, the grid cell is liquid when the phase fraction alpha is 1, the grid cell is filled with air when the phase fraction alpha is 0, and the grid cell exists at the free liquid level when the phase fraction alpha is between 0 and 1. The side of the numerical wave pool is a wall type boundary to prevent water from flowing out. The top is an atm sphere boundary that prevents backflow phenomenon for equalizing the pressure in the pool. The boundary conditions of the wave pool with specific values are shown in the following table:
step three: generation of motion float
The numerical pool can simulate the motion condition of objects with various shapes in waves, the simulated objects can be generated by 3D drawing software and then imported into the numerical pool, grids attached to the surfaces of the objects are generated around the objects through a self-contained tool snappyHexmesh of OpenFOAM, and the grid density near the objects can be set in a control file. Taking a "convex" floating body as an example, the grid is shown in fig. 3.
Step four: wave generation
The wave generation method of the numerical wave pool is provided by a waves2Foam library, and specific liquid speeds are set at the wave making boundary to generate waves. Taking Stokes second-order rule waves as an example, the wave speed distribution formula is as follows:
where u (x, z, t) and w (x, z, t) are the horizontal and vertical components, respectively, of the wave velocity. H is the wave height of the wave, ω is the angular frequency of the wave, and H is the water depth. In Stokes regular waves, the water depth, the wave period and the wave height are parameters which can be set, and different parameters correspond to different waves.
Step five: solving by wave-object interaction solver
The invention integrates the solver wave Foam in the two-phase flow solver InterDyMFOAM and the waves2Foam of the OpenFOAM, and develops a solver suitable for interaction of waves and moving objects. Newly developed solvers can solve two-phase flow problems under wave conditions and can solve object motion using deformed mesh techniques.
The specific solving process is as follows: and (3) obtaining a continuity equation, a momentum equation and an energy equation through a physical balance law, and integrating on each control body to obtain an algebraic equation set. Then solving an equation set to obtain flow field variables of each grid, obtaining stress data of the object through flow field information, solving motion variables of the object such as linear velocity, displacement, angular velocity and the like through a Newmark method, and updating positions of grid points of the movable grid to realize motion of the object.
The second part of the invention is a neural network based on the GRU structure that predicts the forces of the floating body in the waves.
The invention establishes a double-layer circulating neural network by utilizing GRU units for predicting the stress of the floating body in the wave in real time. The network structure is shown in fig. 5. Every time step, the neural network receives a group of load gesture data, and the two layers of GRU units expand input information to a high-dimensional space for processing. The GRU is a recurrent neural network, and the hidden information of the previous input data is recorded through the memory unit h. And finally, the data passes through a full connection layer to output stress information of the object. The input of each time step of the neural network is a one-dimensional vector with 9 elements, and the one-dimensional vector comprises wave parameters and attitude information of an object. The output is a one-dimensional vector of magnitude 6, which is the force and moment experienced by the load in the x, y, z-axis directions, respectively.
The training data of the neural network are obtained by using the numerical wave pool mentioned in the embodiment example through computational fluid dynamics simulation, and the motion response of the floating body under different sea conditions is simulated, wherein different wave conditions are realized by changing wave height, wave period and water depth in the numerical wave pool. Each set of floating bodies interacted with the wave simulates a time of 20s in a numerical wave pool, the maximum time step during the simulation is 0.0001s, the time step is set to be automatically adjustable, and the maximum kurroa number is 0.8. All data obtained by the cases are vertically spliced into a training set during each training, the splicing sequence of each training set is randomly disordered, and finally a training set with the number of about 10 ten thousand data is obtained, wherein parameters of the training set are shown in table 1.
During training, the loss function of the model is the Mean Square Error (MSE), and the expression is:
wherein n is the data amount, y i Is a predicted value of the current value,is a reference value obtained by a numerical wave pool. Using Adam algorithm as a gradient descent method, the learning rate was set to 0.00005, and the dropout coefficient was set to 0.1 in order to avoid overfitting. The whole training process circularly trains the training set for 1200 times.
Table 1 training set data parameters
The neural network after final training is used for predicting the stress of the floating body in the wave, the variance result of the neural network on the training set is shown in fig. 6, and the horizontal force F can be obtained x The predicted result of (2) is worst, vertical force F z The predicted outcome of (2) is best. As shown in fig. 7-9, it can be seen that the predicted data for the trained neural network is substantially the same as the data obtained for the computational fluid dynamics approach. The neural network predicts that the calculation time of the stress data of the object is very short, and almost predicts the stress information of the object in the waves in real time.
In order to prove that the trained neural network can better realize the function, the prediction precision of the neural network is tested by using data which is not in a training set, parameters of the testing set are shown in table 2, mean square values of the testing set are shown in fig. 10, and partial data comparison results are shown in fig. 11-12. It can be seen that while the predictive effect is not as good as the test set, the overall is still satisfactory.
Table 2 test set data parameters
On the basis, the trained neural network is combined with a numerical wave pool:
the PyToch library is introduced into the computational fluid dynamics software OpenFOAM, a trained network is added into a numerical wave pool, corresponding acting force is applied to the floating body through designing a controller, the feedback input of the controller is the numerical value predicted by the neural network, the movement of the controlled floating body is recorded, and the effectiveness of the controller is verified, wherein a schematic diagram is shown in FIG. 13.
Wave and object interaction simulation example
Taking the interaction of the floating body in the shape of a Chinese character 'yang' and waves as an example, the motion and stress of the floating body in the regular waves are solved, as shown in fig. 4 (a). At the beginning of the simulation, the float was placed in the middle of the numeric wave pool and the float was constrained from horizontal displacement. The wave starts to generate after the simulation starts, the floating body starts to move along with the wave, and the heave and rotation data of the floating body are shown in fig. 4 (b). The solid line is the physical experiment result, the broken line is the simulation data result of the numerical wave pool, and the results of the two groups of object movement can be seen to be basically consistent. The accuracy of the numerical wave pool calculation of the wave-object interaction problem is demonstrated.
The invention designs a numerical wave pool for solving the interaction of a moving object and waves by using an OpenFOAM and waves2Foam library, and develops a new solver to solve the problem of a moving grid.
The invention designs a neural network based on GRU units to predict wave force applied to a moving object, and training data of the neural network are obtained through the numerical wave pool.
According to the invention, a controller can be designed in the numerical wave pool, a controller interface is preset in the numerical wave pool, namely, a control equation can be added in a configuration file, in the simulation calculation process, the control force is obtained through feedback information calculation and then applied to a controlled object, and the object is controlled to stabilize the object motion. The feedback information can be obtained through simulation calculation or through a neural network.
According to the invention, the combination of the neural network and the computational fluid dynamics software OpenFOAM is realized by importing the C++ code library of PyTorch. A trained god is added into the numerical wave pool, and the designed controller can control the object by utilizing the data predicted by the neural network, so that the feasibility of the controller combined with the neural network is verified.
Example two
It is an object of the present embodiment to provide a computing device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, which processor implements the steps of the method described above when executing the program.
Example III
An object of the present embodiment is to provide a computer-readable storage medium.
A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of the above method.
Example IV
It is an object of this embodiment to provide a real-time prediction system of the wave forces to which a moving load is subjected, comprising:
a simulation model building module configured to: establishing a numerical wave pool simulating the interaction of the simulated wave and the floating body and a solver solving the interaction of the wave and the floating body;
a solution module configured to: the interaction of the floating bodies in two-dimensional and three-dimensional different shapes and the waves with different parameters is solved by using the solver, and the motion gesture of the floating bodies along with the waves and the stress information at each moment are calculated;
an online prediction module configured to: and training the neural network by using the calculated motion gesture and stress information, and predicting the wave force of the floating body in a specific environment on line by using the trained neural network.
The steps involved in the devices of the second, third and fourth embodiments correspond to those of the first embodiment of the method, and the detailed description of the embodiments can be found in the related description section of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present invention.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, whereby they may be stored in storage means for execution by computing means, or they may be made into individual integrated circuit modules separately, or a plurality of modules or steps in them may be made into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.

Claims (7)

1. The real-time prediction method for the wave force of the motion load is characterized by comprising the following steps:
establishing a numerical wave pool simulating the interaction of the simulated wave and the floating body and a solver solving the interaction of the wave and the floating body;
the numerical wave pool grid and boundary condition setting are required to be set before the numerical wave pool simulates interaction between the simulated wave and the floating body;
the set numerical wave pool grid is specifically:
setting the size of a numerical wave pool;
integrating each grid by utilizing the hexahedral grid discrete calculation region and obtaining a discrete equation set of the whole calculation region;
when a hexahedral grid is used for discretizing a calculation area, the discrete grids are set to be non-uniform or uniform-size grids, and the variation degree of a flow field variable in the calculation area is positively related to the density degree of the grids;
in the simulation of the interaction of waves and objects, a wave generation boundary, namely a region on the left side of a pool, a region where a floating body is located, namely a region in the middle of the pool, and a wave absorption region, namely a region on the right side of the pool, adopt grids with higher density;
when setting boundary conditions, the constrained variables include velocity, pressure, and free surface;
the numerical wave pool adopts a fluid volume method to solve the shape of the free liquid level, the phase fraction alpha is used for distinguishing air from liquid, the grid cell is liquid when the phase fraction alpha is 1, the grid cell is filled with air when the phase fraction alpha is 0, and the grid cell exists at the free liquid level when the phase fraction alpha is between 0 and 1;
the side surface of the numerical wave pool is a wall type boundary, water is prevented from flowing out, the top is an atm boundary, and backflow is prevented, so that the pressure in the pool is balanced;
the interaction of the floating bodies in two-dimensional and three-dimensional different shapes and the waves with different parameters is solved by using the solver, and the motion gesture of the floating bodies along with the waves and the stress information at each moment are calculated;
training a neural network by using the calculated motion gesture and stress information, and predicting the wave force of the floating body in a specific environment on line by using the trained neural network;
the trained neural network predicts the wave force of the floating body in a specific environment on line, and specifically comprises the following steps:
each time step, the neural network receives a group of load gesture data, and the two layers of GRU units expand input information to a high-dimensional space for processing;
the GRU is a cyclic neural network, and hidden information of the input data is recorded through the memory unit h;
finally, the data passes through a full connection layer to output stress information of the object;
the input of each time step of the neural network is a one-dimensional vector with a plurality of elements, the one-dimensional vector respectively comprises wave parameters and attitude information of an object, and the output is a one-dimensional vector which is respectively the force and moment born by the floating body in the directions of x, y and z axes.
2. The method for predicting the motion load under the wave force in real time according to claim 1, wherein the numerical wave pool is a three-dimensional simulation model, an origin o is placed at the bottom of the numerical wave pool, an x-axis is horizontally right in a direction parallel to the still water surface, a z-axis is vertically upward, a y-axis is a direction in which a width edge of the numerical wave pool is located, and waves propagate along the x-axis.
3. The method for predicting the motion load in real time under the wave force according to claim 1, wherein when a hexahedral grid is used for discrete calculation of the area, the discrete grid is set to be a non-uniform or uniform-sized grid, and the variation degree of the flow field variable in the calculation area is positively correlated with the density degree of the grid;
in the simulation of wave and object interactions, the wave generation boundary, i.e., the left side region of the pool, the floating body region, i.e., the middle region of the pool, and the wave absorption region, i.e., the right side region of the pool, employ higher density grids.
4. A method for real-time prediction of the wave force of a moving load according to claim 1, wherein the floating body is generated and then introduced into a numerical wave pool, and a desired grid is generated around the floating body.
5. The real-time prediction simulation system for the wave force of the motion load is characterized by comprising:
a simulation model building module configured to: establishing a numerical wave pool simulating the interaction of the simulated wave and the floating body and a solver solving the interaction of the wave and the floating body;
the numerical wave pool grid and boundary condition setting are required to be set before the numerical wave pool simulates interaction between the simulated wave and the floating body;
the set numerical wave pool grid is specifically:
setting the size of a numerical wave pool;
integrating each grid by utilizing the hexahedral grid discrete calculation region and obtaining a discrete equation set of the whole calculation region;
when a hexahedral grid is used for discretizing a calculation area, the discrete grids are set to be non-uniform or uniform-size grids, and the variation degree of a flow field variable in the calculation area is positively related to the density degree of the grids;
in the simulation of the interaction of waves and objects, a wave generation boundary, namely a region on the left side of a pool, a region where a floating body is located, namely a region in the middle of the pool, and a wave absorption region, namely a region on the right side of the pool, adopt grids with higher density;
when setting boundary conditions, the constrained variables include velocity, pressure, and free surface;
the numerical wave pool adopts a fluid volume method to solve the shape of the free liquid level, the phase fraction alpha is used for distinguishing air from liquid, the grid cell is liquid when the phase fraction alpha is 1, the grid cell is filled with air when the phase fraction alpha is 0, and the grid cell exists at the free liquid level when the phase fraction alpha is between 0 and 1;
the side surface of the numerical wave pool is a wall type boundary, water is prevented from flowing out, the top is an atm boundary, and backflow is prevented, so that the pressure in the pool is balanced;
a solution module configured to: the interaction of the floating bodies in two-dimensional and three-dimensional different shapes and the waves with different parameters is solved by using the solver, and the motion gesture of the floating bodies along with the waves and the stress information at each moment are calculated;
an online prediction module configured to: training a neural network by using the calculated motion gesture and stress information, and predicting the wave force of the floating body in a specific environment on line by using the trained neural network;
the trained neural network predicts the wave force of the floating body in a specific environment on line, and specifically comprises the following steps:
each time step, the neural network receives a group of load gesture data, and the two layers of GRU units expand input information to a high-dimensional space for processing;
the GRU is a cyclic neural network, and hidden information of the input data is recorded through the memory unit h;
finally, the data passes through a full connection layer to output stress information of the object;
the input of each time step of the neural network is a one-dimensional vector with a plurality of elements, the one-dimensional vector respectively comprises wave parameters and attitude information of an object, and the output is a one-dimensional vector which is respectively the force and moment born by the floating body in the directions of x, y and z axes.
6. A computing device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method of any of the preceding claims 1-4 when the program is executed.
7. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, performs the steps of the method of any of the preceding claims 1-4.
CN202110943138.5A 2021-08-17 2021-08-17 Real-time prediction method and simulation system for wave force applied to motion load Active CN113609782B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110943138.5A CN113609782B (en) 2021-08-17 2021-08-17 Real-time prediction method and simulation system for wave force applied to motion load

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110943138.5A CN113609782B (en) 2021-08-17 2021-08-17 Real-time prediction method and simulation system for wave force applied to motion load

Publications (2)

Publication Number Publication Date
CN113609782A CN113609782A (en) 2021-11-05
CN113609782B true CN113609782B (en) 2023-08-22

Family

ID=78340954

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110943138.5A Active CN113609782B (en) 2021-08-17 2021-08-17 Real-time prediction method and simulation system for wave force applied to motion load

Country Status (1)

Country Link
CN (1) CN113609782B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114492220B (en) * 2021-11-23 2022-09-16 华北电力大学 OpenFOAM platform-based method and system for predicting flow characteristics of ocean atmospheric boundary layer
CN114510887B (en) * 2022-01-20 2022-11-15 哈尔滨工程大学 Method for rapidly predicting liquid level of large container under ocean condition based on proxy model
CN114624002B (en) * 2022-03-15 2023-02-14 浙江大学 Intelligent wave generation method based on machine learning
CN114715331B (en) * 2022-06-02 2022-09-09 中国海洋大学 Floating ocean platform power positioning control method and system
CN116011356B (en) * 2023-01-09 2023-09-08 浙江大学 Submarine pipeline scouring prediction method based on ConvLSTM and OpenFOAM numerical computation coupling

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111506969A (en) * 2020-04-21 2020-08-07 常熟理工学院 Ship type optimization method based on multi-target particle swarm algorithm
CN111506968A (en) * 2020-04-21 2020-08-07 常熟理工学院 Ship type optimization method based on BP neural network algorithm
CN111506970A (en) * 2020-04-21 2020-08-07 常熟理工学院 Ship hydrodynamic performance evaluation method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111506969A (en) * 2020-04-21 2020-08-07 常熟理工学院 Ship type optimization method based on multi-target particle swarm algorithm
CN111506968A (en) * 2020-04-21 2020-08-07 常熟理工学院 Ship type optimization method based on BP neural network algorithm
CN111506970A (en) * 2020-04-21 2020-08-07 常熟理工学院 Ship hydrodynamic performance evaluation method

Also Published As

Publication number Publication date
CN113609782A (en) 2021-11-05

Similar Documents

Publication Publication Date Title
CN113609782B (en) Real-time prediction method and simulation system for wave force applied to motion load
Meneghini et al. Numerical simulation of flow interference between two circular cylinders in tandem and side-by-side arrangements
CN105260581B (en) The virtual vibration and shock testing method of naval vessel electromechanical control equipment
Wei et al. Chaos in breaking waves
Lu et al. Multi-scale modeling of shock interaction with a cloud of particles using an artificial neural network for model representation
Sohn et al. Numerical analysis of vortex core phenomenon during draining from cylinder tank for various initial swirling speeds and various tank and drain port sizes
Jena et al. A numerical study of violent sloshing problems with modified MPS method
CN110598168B (en) Numerical simulation method and system for object entering water
Lakshmynarayanana et al. Coupled fluid structure interaction to model three-dimensional dynamic behaviour of ship in waves
Pinkster et al. A real-time simulation technique for ship-ship and ship-port interactions
Aksenov et al. Ship hull form design and optimization based on CFD
CN113312857B (en) Simulation method of square waves in numerical value pool based on SPH
CN114218672A (en) Buffeting response obtaining method for cross-supersonic-velocity large-attack-angle trim wing and related device
Mnasri et al. A moving grid model for simulation of free surface behavior induced by horizontal cylinders exit and entry
Makarov et al. Parametric sensitivity of crane ship numerical model with experimental verification in a wave basin
Johannessen Use of CFD to Study Hydrodynamic Loads on Free-Fall Lifeboats in the Impact Phase.: A verification and validation study.
CN113673007B (en) SPH-based method for forecasting wave resistance of ship in bidirectional waves
Eskilsson et al. Estimation of nonlinear forces acting on floating bodies using machine learning
López et al. Framework for the shape optimization of aerodynamic profiles using genetic algorithms
Mola et al. Ship sinkage and trim predictions based on a CAD interfaced fully nonlinear potential model
Zhang et al. Fluid-structure interaction problem of hydrodynamic motion of an independent escape capsule under variable constraint conditions
Kudupudi et al. Numerical investigation of the effect due to vessel motion on green water impact on deck
Skorpa Numerical simulation of flow around remotely operated vehicle (ROV)
Paci et al. 3D-numerical analysis of wave-floating structure interaction with OpenFOAM
Jeong et al. Numerical simulations of two-dimensional floating breakwaters in regular waves using fixed cartesian grid

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant